feat(orders): competitiveness report for your open orders

[piekstra]

What

New read-only command lofty orders competitiveness [--property-id ID] (alias comp; --json → orders-competitiveness/v1). For each property you have open orders on, per side:

Side	Fill gap (dollars and %)	Competition gap (dollars and %)
buy	your highest bid up to the lowest ask not yours, as % of that ask (how far a seller must come down to fill you)	book's highest bid vs the highest bid not yours, as % of the latter
sell	highest bid not yours up to your lowest ask, as % of that bid	lowest ask not yours vs the book's lowest ask, as % of the former

Plus leadUsd (signed distance ahead of the best competing order) and atTop.

Design choices

• Counterparty excludes your own orders. When you quote both sides, your own ask is often the book's best ask, and it cannot fill your bid. Measuring the fill gap against it would just report your own spread. The book's raw best bid and ask are still reported in the book block.
• Attributing book quantity. /properties/{id}/orderbook is aggregated by price with no order ids. Your share of a level is your open quantity at that price, and the remainder belongs to someone else. A shared top level therefore has a competition gap of 0. The legacy per-order envelope (orderBook.buyOrders) is matched by order id.
• Partial fills. Open means status: active. A partially filled order stays active at its remaining quantity in the all=true list.

Also fixed: orders list --property-id

orders list --property-id used the ?propertyId= query, which reports a partially filled order's original quantity. The all=true list and the single-order GET report what remains (quote already works around this). The command now filters the all=true list client-side, and the response shape is unchanged. The new command uses the same path. Live check: the order set for one property is identical under both queries (no partial fills were open to compare quantities).

Shared book and open-order rules

• src/commands/book.rs now owns order-book parsing for both envelopes and the "book without my own orders" rule. The quote never-cross rail, orders create's market check, rewards, and this command all read through it, so the rail and the report cannot disagree. Book levels with no positive quantity no longer count as resting liquidity anywhere.
• orders::open_orders is the only place that fetches open orders (all=true, filtered client-side). quote, rewards eligibility, account coverage, and this command use it.

Tests

• 5 unit tests in src/commands/book.rs (both envelopes, quantity subtraction, id matching, side isolation, empty side), 1 for retain_property, and 8 in src/commands/competitiveness.rs over synthetic fixtures in tests/fixtures/competitiveness/: shared top level, alone on top, undercut ask (negative lead), legacy id envelope, empty opposite side, one-sided orders, grouping by property.
• Shape test: the synthetic orders and book levels carry every field of the captured live fixtures.
• Surface test: orders --help lists it and orders comp --help renders.
• Existing quote and rewards tests pass unchanged on the shared module.
• make verify green locally. Live read-only runs of orders comp, account coverage, and orders list --property-id exit 0. rewards eligibility exits 4 (no open orders), the same result the released binary gives. Live read-only run: exit 0, orders-competitiveness/v1 emitted (the account had no open orders at the time, so the per-property path was covered by fixtures). The live order book shape (orderbook.{bids,asks}, fractional quantities) matches what the parser reads.

No orders are placed or cancelled.

[piekstra-dev]

Automated PR Review

Reviewed commit: 560182b5f213
Profile: reviewer - Posting as: piekstra-dev

Summary

Reviewer	Findings
rust:implementation-tests	0
policies:conventions	0
architecture:solid	3
documentation:docs	0
rust:implementation	1

architecture:solid (3 findings)

Major - src/commands/competitiveness.rs:41

U-S1 / U-O1: side() is a second, divergent implementation of the domain rule "the book with my own resting size removed". quote::market_from (src/commands/quote.rs:438-490) already computes exactly this; it backs the never-cross-the-market rail. The two implementations now disagree on mechanics. Price matching is cents() rounding here and abs() < 0.005 there. The residual threshold is > 1e-9 here and > 0.0 there. On the legacy envelope, this module matches by order id/orderId, while quote subtracts by price from the first matching entry. The PR's design note relies on that id matching. The result is that orders competitiveness and quote can report different values for "best ask not yours" on the same book, and the rail is the one that moves money. This diff also adds the 6th copy of the <path>/{bids,asks} vs <path>/{buyOrders,sellOrders} envelope fallback; the others are properties.rs:112, quote.rs:442, rewards.rs:238 and :374, and orders.rs:348. A third envelope shape would mean editing 6 sites. Suggested fix: add one pure module (for example src/commands/book.rs) exposing fn levels(book: &Value, side: BookSide) -> Vec<Level { cents: i64, qty: f64, id: Option<String> }> and fn without_mine(levels: &[Level], mine: &[&Value]) -> Vec<Level>. Carry the id-aware attribution over from this PR. Have both competitiveness::side and quote::market_from consume it, and move the fixture-driven tests that live here onto it. Migrating rewards and properties can follow separately. If the two consumers are meant to differ, document why at both sites.

Minor - src/commands/orders.rs:117

U-S1: the rule "open orders come from the all=true list, filtered client-side to status == active (and optionally to one property), because ?propertyId= reports original quantity after a partial fill" is now hand-copied in orders.rs:86-93, orders.rs:117-128, quote.rs:289-300, rewards.rs:168-174 and account.rs:178/207. This PR also writes the rule into AGENTS.md as prose. A manifest rule kept in 5+ copies by convention only takes one new command that uses ?propertyId= to break it. Suggested fix: add fn open_orders(client: &LoftyClient, property_id: Option<&str>) -> Result<Vec<Value>, CliError> next to retain_property, or in commands/mod.rs, and call it from Competitiveness and quote::fetch_state. Point the AGENTS.md sentence at that function so the rule has one owner.

Minor - src/commands/orders.rs:250

U-T1: orders list --property-id changed behavior. It used to be the server-side ?propertyId= query and is now all=true plus client-side retain_property. Nothing in the diff tests it: the surface tests only check that --property-id parses. retain_property is pure and private, and other top-level fields are documented as kept unchanged, so it is cheap to pin. Suggested fix: add a unit test in the existing orders.rs mod tests that runs retain_property on a two-property <path> payload. It should assert that only the target property's orders remain and that a sibling top-level field is untouched. Reuse tests/fixtures/competitiveness/orders-open.json, which already holds two properties.

rust:implementation (1 finding)

Minor - src/commands/orders.rs:8

market_view's side closure in orders.rs re-implements the same orderbook-level lookup (<path>/{bids,asks} falling back to the legacy <path>/{buyOrders,sellOrders} envelope) that competitiveness::side already implements independently. A third book shape or a fix to the fallback logic now has to be made in two places, and they have already drifted once (competitiveness::side filters quantity <= 0 per level; market_view::side does not, since it only tracks min/max price). Factor the shared "find the levels array for this side" step into one function both call.

Reviewer Coverage

• rust:implementation-tests — complete (broad); inspected 5 assigned files (13 inspected across reviewers): src/commands/competitiveness.rs, src/commands/mod.rs, src/commands/orders.rs, tests/cli_surface.rs, tests/fixture_shapes.rs; skipped: none; constraints: none
• policies:conventions — complete (broad); inspected 5 assigned files (13 inspected across reviewers): AGENTS.md, README.md, src/commands/competitiveness.rs, src/commands/mod.rs, src/commands/orders.rs; skipped: none; constraints: Base-branch .codereview/agents/ guidance was not present for this review (dossier reports it missing). No ../cli-common or ../.github sibling convenience copies present in the workbench; relied on this repo's AGENTS.md as the source of truth for conventions.
• architecture:solid — complete (broad); inspected 7 assigned files (13 inspected across reviewers): src/commands/competitiveness.rs, src/commands/orders.rs, tests/fixtures/competitiveness/orderbook-alone-on-top.json, tests/fixtures/competitiveness/orderbook-bids-only.json, tests/fixtures/competitiveness/orderbook-legacy-ids.json, tests/fixtures/competitiveness/orderbook-shared-top.json, tests/fixtures/competitiveness/orders-open.json; skipped: none; constraints: AGENTS.md was read as the project manifest. It contains no rule IDs, so findings cite the universal U-* rules. git and cargo commands required interactive approval in this environment, so the base-vs-head diff was reconstructed from head files plus the change map, and make verify / cargo test were not run; build and test status is taken from the PR description only.
• documentation:docs — complete (broad); inspected 3 assigned files (13 inspected across reviewers): AGENTS.md, README.md, tests/fixtures/README.md; skipped: none; constraints: none
• rust:implementation — complete (broad); inspected 3 assigned files (13 inspected across reviewers): src/commands/competitiveness.rs, src/commands/mod.rs, src/commands/orders.rs; skipped: none; constraints: none

Inspected files (13)

• AGENTS.md
• README.md
• src/commands/competitiveness.rs
• src/commands/mod.rs
• src/commands/orders.rs
• tests/cli_surface.rs
• tests/fixture_shapes.rs
• tests/fixtures/README.md
• tests/fixtures/competitiveness/orderbook-alone-on-top.json
• tests/fixtures/competitiveness/orderbook-bids-only.json
• tests/fixtures/competitiveness/orderbook-legacy-ids.json
• tests/fixtures/competitiveness/orderbook-shared-top.json
• tests/fixtures/competitiveness/orders-open.json

0 PR discussion threads considered. 0 summarized; 0 resolved.

---

Completed in 4m 21s | $2.55 | claude-sonnet-5, claude-opus-5-5 | cr 0.10.314

Field	Value
Model	claude-sonnet-5, claude-opus-5-5
Reviewers	rust:implementation-tests, policies:conventions, architecture:solid, documentation:docs, rust:implementation
Engine	claude_cli · claude-sonnet-5, claude-opus-5-5
Reviewed by	cr · piekstra-dev
Duration	4m 21s wall · 7m 51s compute
Cost	$2.55
Tokens	122 in / 36.8k out

Per-workstream usage

• orchestrator-selection — claude-sonnet-5
  • In: 6
  • Out: 3.5k
  • Cache read: 138.6k
  • Cache create: 18.8k
  • Cost: $0.14
  • Duration: 41s
• rust:implementation-tests — claude-sonnet-5
  • In: 20
  • Out: 7.6k
  • Cache read: 670.1k
  • Cache create: 42.5k
  • Cost: $0.38
  • Duration: 1m 42s
• policies:conventions — claude-sonnet-5
  • In: 22
  • Out: 3.1k
  • Cache read: 620.2k
  • Cache create: 23.2k
  • Cost: $0.25
  • Duration: 37s
• architecture:solid — claude-opus-5-5
  • In: 28
  • Out: 8.5k
  • Cache read: 936.0k
  • Cache create: 56.1k
  • Cost: $0.81
  • Duration: 1m 28s
• documentation:docs — claude-sonnet-5
  • In: 28
  • Out: 4.9k
  • Cache read: 869.9k
  • Cache create: 36.4k
  • Cost: $0.37
  • Duration: 1m 23s
• rust:implementation — claude-sonnet-5
  • In: 12
  • Out: 8.6k
  • Cache read: 354.9k
  • Cache create: 42.2k
  • Cost: $0.33
  • Duration: 1m 36s
• orchestrator-rollup — claude-sonnet-5
  • In: 6
  • Out: 592
  • Cache read: 201.4k
  • Cache create: 6.3k
  • Cost: $0.28
  • Duration: 22s

[piekstra-dev]

Automated PR Review

Reviewed commit: 9c77cc48e052
Profile: reviewer - Posting as: piekstra-dev

Summary

Reviewer	Findings
rust:implementation-tests	0
policies:conventions	0
architecture:solid	1
documentation:docs	0
rust:implementation	0

architecture:solid (1 finding)

Minor - src/commands/orders.rs:362

U-T1: the orders create market check changes behavior here. The old side closure counted any level that had a numeric price. book::levels now also requires a numeric quantity greater than 1e-9 (book.rs:67 and :73). A zero-quantity level no longer counts; the PR describes that as intended. A level with no quantity field is now silently dropped as well, and the PR does not mention that case. For this check, dropping a level fails open: an ask the book does report becomes invisible to the 'AT OR ABOVE the best ask' objection. AGENTS.md says changes to money-moving checks are safety-critical, but no test pins either case. The book() helper in orders.rs tests always sets quantity 1, and book.rs tests use only positive-quantity fixtures. Risk today is low: tests/fixture_shapes.rs shows live levels carry quantity. The verdict would change to no finding if a test pinned both cases. Suggested fix: add a test to the orders.rs mod tests where market_view gets a book whose best ask has quantity 0 and whose next ask is real, then assert the objection quotes the real ask. Decide explicitly how a level without quantity should be treated. Failing closed means counting it as resting for best-price purposes, for example qty.unwrap_or(f64::INFINITY) or a separate prices() reader. Pin whichever choice is made in a book.rs test.

Reviewer Coverage

• rust:implementation-tests — complete (constrained); inspected 9 assigned files (17 inspected across reviewers): src/commands/account.rs, src/commands/book.rs, src/commands/competitiveness.rs, src/commands/mod.rs, src/commands/orders.rs, src/commands/quote.rs, src/commands/rewards.rs, tests/cli_surface.rs, tests/fixture_shapes.rs; skipped: none; constraints: none
• policies:conventions — complete (constrained); inspected 5 assigned files (17 inspected across reviewers): AGENTS.md, README.md, src/commands/competitiveness.rs, src/commands/mod.rs, src/commands/orders.rs; skipped: none; constraints: Base-branch .codereview/agents/ guidance was not present for this review (dossier reports it missing). No ../cli-common or ../.github sibling convenience copies present in the workbench; relied on this repo's AGENTS.md as the source of truth for conventions. src/commands/book.rs, account.rs, quote.rs, and rewards.rs are not in this reviewer's assigned file list; verified their consumption of the shared book module only to confirm the AGENTS.md claims in the assigned files, not reviewed line-by-line.
• architecture:solid — complete (constrained); inspected 7 assigned files (17 inspected across reviewers): src/commands/competitiveness.rs, src/commands/orders.rs, tests/fixtures/competitiveness/orderbook-alone-on-top.json, tests/fixtures/competitiveness/orderbook-bids-only.json, tests/fixtures/competitiveness/orderbook-legacy-ids.json, tests/fixtures/competitiveness/orderbook-shared-top.json, tests/fixtures/competitiveness/orders-open.json; skipped: none; constraints: All four earlier threads are resolved at this head: book.rs owns envelope parsing and own-size removal, open_orders owns the fetch rule, market_view uses book::levels, and retain_property has a unit test (orders.rs:546). git and cargo commands required interactive approval in this environment, so make verify / cargo test were not run and the diff was judged from head files plus the change map; build and test status is taken from the PR description. src/commands/book.rs, quote.rs, rewards.rs and account.rs were read for context only; they are outside this reviewer's assigned files, so findings anchor on orders.rs.
• documentation:docs — complete (constrained); inspected 3 assigned files (17 inspected across reviewers): AGENTS.md, README.md, tests/fixtures/README.md; skipped: none; constraints: none
• rust:implementation — complete (constrained); inspected 3 assigned files (17 inspected across reviewers): src/commands/competitiveness.rs, src/commands/mod.rs, src/commands/orders.rs; skipped: none; constraints: book.rs is the new home of the previously-duplicated orderbook parsing logic but is not in this reviewer's assigned file list, so it was read for context only; no findings were filed against it.

Inspected files (17)

• AGENTS.md
• README.md
• src/commands/account.rs
• src/commands/book.rs
• src/commands/competitiveness.rs
• src/commands/mod.rs
• src/commands/orders.rs
• src/commands/quote.rs
• src/commands/rewards.rs
• tests/cli_surface.rs
• tests/fixture_shapes.rs
• tests/fixtures/README.md
• tests/fixtures/competitiveness/orderbook-alone-on-top.json
• tests/fixtures/competitiveness/orderbook-bids-only.json
• tests/fixtures/competitiveness/orderbook-legacy-ids.json
• tests/fixtures/competitiveness/orderbook-shared-top.json
• tests/fixtures/competitiveness/orders-open.json

0 PR discussion threads considered. 0 summarized; 0 resolved.

---

Completed in 1m 26s | $4.37 | claude-sonnet-5, claude-opus-5-5 | cr 0.10.314

Field	Value
Model	claude-sonnet-5, claude-opus-5-5
Reviewers	rust:implementation-tests, policies:conventions, architecture:solid, documentation:docs, rust:implementation
Engine	claude_cli · claude-sonnet-5, claude-opus-5-5
Reviewed by	cr · piekstra-dev
Duration	1m 26s wall · 4m 20s compute
Cost	$4.37
Tokens	94 in / 18.4k out

Per-workstream usage

• rust:implementation-tests — claude-sonnet-5
  • In: 22
  • Out: 3.7k
  • Cache read: 1.2M
  • Cache create: 44.4k
  • Cost: $0.84
  • Duration: 1m 02s
• policies:conventions — claude-sonnet-5
  • In: 20
  • Out: 2.6k
  • Cache read: 758.2k
  • Cache create: 13.7k
  • Cost: $0.48
  • Duration: 29s
• architecture:solid — claude-opus-5-5
  • In: 14
  • Out: 5.2k
  • Cache read: 789.7k
  • Cache create: 40.8k
  • Cost: $1.39
  • Duration: 55s
• documentation:docs — claude-sonnet-5
  • In: 20
  • Out: 3.4k
  • Cache read: 935.4k
  • Cache create: 21.8k
  • Cost: $0.68
  • Duration: 1m 04s
• rust:implementation — claude-sonnet-5
  • In: 12
  • Out: 3.1k
  • Cache read: 599.9k
  • Cache create: 38.5k
  • Cost: $0.63
  • Duration: 37s
• orchestrator-rollup — claude-sonnet-5
  • In: 6
  • Out: 419
  • Cache read: 219.2k
  • Cache create: 5.0k
  • Cost: $0.35
  • Duration: 10s

[piekstra-dev]

Automated PR Review

Reviewed commit: 92660a442646
Profile: reviewer - Posting as: piekstra-dev

Summary

Reviewer	Findings
rust:implementation-tests	0
policies:conventions	0
architecture:solid	0
documentation:docs	0
rust:implementation	0

Reviewer Coverage

• rust:implementation-tests — complete (constrained); inspected 9 assigned files (18 inspected across reviewers): src/commands/account.rs, src/commands/book.rs, src/commands/competitiveness.rs, src/commands/mod.rs, src/commands/orders.rs, src/commands/quote.rs, src/commands/rewards.rs, tests/cli_surface.rs, tests/fixture_shapes.rs; skipped: none; constraints: none
• policies:conventions — complete (constrained); inspected 5 assigned files (18 inspected across reviewers): AGENTS.md, README.md, src/commands/competitiveness.rs, src/commands/mod.rs, src/commands/orders.rs; skipped: none; constraints: Base-branch .codereview/agents/ guidance was not present for this review (dossier reports it missing). No ../cli-common or ../.github sibling convenience copies present in the workbench; relied on this repo's AGENTS.md as the source of truth for conventions. src/commands/book.rs and rewards.rs are not in this reviewer's assigned file list; checked only to confirm the fail-closed fix and its test coverage, not reviewed line-by-line.
• architecture:solid — complete (constrained); inspected 8 assigned files (18 inspected across reviewers): src/commands/competitiveness.rs, src/commands/orders.rs, tests/fixtures/book/orderbook-zero-and-missing-quantity.json, tests/fixtures/competitiveness/orderbook-alone-on-top.json, tests/fixtures/competitiveness/orderbook-bids-only.json, tests/fixtures/competitiveness/orderbook-legacy-ids.json, tests/fixtures/competitiveness/orderbook-shared-top.json, tests/fixtures/competitiveness/orders-open.json; skipped: none; constraints: All five earlier threads are resolved at this head. The orders.rs:362 thread is closed by book.rs:72-75 (no quantity field means infinite size, failing closed), the orders.rs:577 market_view test, and the book.rs:222 test, both over the new fixture. Not reported because book.rs is outside the assigned files: a level with a non-numeric quantity (for example a string) is still dropped, which fails open. Using an infinite quantity as an 'unknown size' marker is enforced only by the book.rs:55-62 doc and an is_finite() check in rewards.rs:353. git and cargo commands required interactive approval in this environment, so make verify / cargo test were not run; build and test status is taken from the PR description.
• documentation:docs — complete (constrained); inspected 3 assigned files (18 inspected across reviewers): AGENTS.md, README.md, tests/fixtures/README.md; skipped: none; constraints: none
• rust:implementation — complete (constrained); inspected 3 assigned files (18 inspected across reviewers): src/commands/competitiveness.rs, src/commands/mod.rs, src/commands/orders.rs; skipped: none; constraints: book.rs is not in this reviewer's assigned file list and was read for context only; no findings were filed against it.

Inspected files (18)

• AGENTS.md
• README.md
• src/commands/account.rs
• src/commands/book.rs
• src/commands/competitiveness.rs
• src/commands/mod.rs
• src/commands/orders.rs
• src/commands/quote.rs
• src/commands/rewards.rs
• tests/cli_surface.rs
• tests/fixture_shapes.rs
• tests/fixtures/README.md
• tests/fixtures/book/orderbook-zero-and-missing-quantity.json
• tests/fixtures/competitiveness/orderbook-alone-on-top.json
• tests/fixtures/competitiveness/orderbook-bids-only.json
• tests/fixtures/competitiveness/orderbook-legacy-ids.json
• tests/fixtures/competitiveness/orderbook-shared-top.json
• tests/fixtures/competitiveness/orders-open.json

0 PR discussion threads considered. 0 summarized; 0 resolved.

---

Completed in 1m 24s | $6.06 | claude-sonnet-5, claude-opus-5-5 | cr 0.10.314

Field	Value
Model	claude-sonnet-5, claude-opus-5-5
Reviewers	rust:implementation-tests, policies:conventions, architecture:solid, documentation:docs, rust:implementation
Engine	claude_cli · claude-sonnet-5, claude-opus-5-5
Reviewed by	cr · piekstra-dev
Duration	1m 24s wall · 3m 31s compute
Cost	$6.06
Tokens	80 in / 14.3k out

Per-workstream usage

• rust:implementation-tests — claude-sonnet-5
  • In: 20
  • Out: 4.2k
  • Cache read: 1.5M
  • Cache create: 29.3k
  • Cost: $1.29
  • Duration: 1m 03s
• policies:conventions — claude-sonnet-5
  • In: 14
  • Out: 1.6k
  • Cache read: 621.1k
  • Cache create: 13.7k
  • Cost: $0.67
  • Duration: 19s
• architecture:solid — claude-opus-5-5
  • In: 10
  • Out: 3.4k
  • Cache read: 699.5k
  • Cache create: 18.4k
  • Cost: $1.75
  • Duration: 35s
• documentation:docs — claude-sonnet-5
  • In: 22
  • Out: 3.2k
  • Cache read: 1.3M
  • Cache create: 22.2k
  • Cost: $1.05
  • Duration: 1m 03s
• rust:implementation — claude-sonnet-5
  • In: 8
  • Out: 1.6k
  • Cache read: 540.8k
  • Cache create: 30.5k
  • Cost: $0.88
  • Duration: 20s
• orchestrator-rollup — claude-sonnet-5
  • In: 6
  • Out: 335
  • Cache read: 233.4k
  • Cache create: 4.2k
  • Cost: $0.41
  • Duration: 9s